Moving Load Identification Of Bridge Based On Dynamic Strain Data

Monitor the moving load of bridge, understand the importance of vehicleparameters on the bridge structurereliability design, operation maintenancemanagement and other aspects;The current moving load identification theory appliedon practical bridge technology is not mature and the current vehicle weight is of highcost and other limitations, thus exploring one kind of simple, fast and effectivemoving load identification method is of great significance.This paper takes advantage of large-scale finite element program ANSYS andAPDL language to establish parameterizedoriginal finite element model of bridgestructure and use bridge static load exploration data to make finite element modelupdating on initial modal. Utilize ANSYS to make dynamic transient analysis ofmoving load on modified finite element model, make computational analysis ondifferent vehicle weight, speed of a motor vehicle and transverse location, and extractresult of strain transducer measure point; Adopt identification method step by step instages to identify lane station and speed firstly and then use computation result toestablish corresponding training sample of neural net of vehicle weight, lane station,vehicle speed and dynamic train peak value at each measure point; adopt BP neuralnet arithmetic to identify moving load vehicle weight by using dynamic train peakdata at each measure point and recognized parameter, thus to respectively establish asimulation model and practical bridge to explore such method; Major research workand achievements:（1）Use spot bridge monitoring and static load data of load experience tomodify finite element model, select the bending stiffness of each component ascorrected parameter, establish proper object function, compare and make research ondifferent optimization method, apply first order search method to get satisfyingamendment;（2）Simulate the moving process of vehicle load on the bridge and consider thedynamic effect influence, based on modified finite element model, use ANSYS andAPDL language program composition to stimulate moving load and make dynamictransient analysis on it; （3）Through comparative analysis, it found the law that the role of the differentlanes T beam peak dynamic strain order of magnitude different, Explore the use ofthis law to identify the lane; Explore a kind of vehicle identification method, firstcalculate the different test dynamic strain difference between the peak section of thecorresponding time point, and calculates the actual distance between the different testsection, and finally dividing by the time difference between the actual distance of thespeed of the vehicle.（4）Have explored the dynamic strain response value under different parametermoving load at each measure point, establish neural net training sample, use thetheory of trained BP neural net to identify vehicle weight;（5）Select the bridge of berth at Guangxi Hengxian County as project exampleto make experiment on the theory of this paper and the experiment result indicates:the bridge moving load identification method based on dynamic strain data is valid.